CN117059198B - Emission list feedback updating method, system and equipment based on response surface model - Google Patents
Emission list feedback updating method, system and equipment based on response surface model Download PDFInfo
- Publication number
- CN117059198B CN117059198B CN202310781360.9A CN202310781360A CN117059198B CN 117059198 B CN117059198 B CN 117059198B CN 202310781360 A CN202310781360 A CN 202310781360A CN 117059198 B CN117059198 B CN 117059198B
- Authority
- CN
- China
- Prior art keywords
- target
- region
- emission
- area
- response surface
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000004044 response Effects 0.000 title claims abstract description 121
- 238000000034 method Methods 0.000 title claims abstract description 49
- 239000003344 environmental pollutant Substances 0.000 claims abstract description 149
- 231100000719 pollutant Toxicity 0.000 claims abstract description 149
- 239000011159 matrix material Substances 0.000 claims abstract description 51
- 238000005070 sampling Methods 0.000 claims abstract description 49
- 238000012544 monitoring process Methods 0.000 claims abstract description 45
- 238000012937 correction Methods 0.000 claims abstract description 28
- 239000002243 precursor Substances 0.000 claims abstract description 22
- 230000008859 change Effects 0.000 claims abstract description 20
- 239000012855 volatile organic compound Substances 0.000 claims description 49
- 238000004088 simulation Methods 0.000 claims description 39
- 230000006870 function Effects 0.000 claims description 35
- 238000004364 calculation method Methods 0.000 claims description 29
- 238000005457 optimization Methods 0.000 claims description 28
- 239000000356 contaminant Substances 0.000 claims description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 13
- 238000010276 construction Methods 0.000 claims description 10
- 230000009467 reduction Effects 0.000 claims description 10
- 238000010845 search algorithm Methods 0.000 claims description 9
- 238000009826 distribution Methods 0.000 claims description 8
- 238000013215 result calculation Methods 0.000 claims description 3
- 239000004065 semiconductor Substances 0.000 claims description 2
- 239000007787 solid Substances 0.000 claims description 2
- 238000004519 manufacturing process Methods 0.000 claims 1
- 239000000809 air pollutant Substances 0.000 abstract description 3
- 231100001243 air pollutant Toxicity 0.000 abstract description 3
- 239000013618 particulate matter Substances 0.000 description 43
- RAHZWNYVWXNFOC-UHFFFAOYSA-N Sulphur dioxide Chemical compound O=S=O RAHZWNYVWXNFOC-UHFFFAOYSA-N 0.000 description 39
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 description 31
- 229910002089 NOx Inorganic materials 0.000 description 31
- 229910000069 nitrogen hydride Inorganic materials 0.000 description 14
- 239000000243 solution Substances 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- JCXJVPUVTGWSNB-UHFFFAOYSA-N nitrogen dioxide Inorganic materials O=[N]=O JCXJVPUVTGWSNB-UHFFFAOYSA-N 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- MGWGWNFMUOTEHG-UHFFFAOYSA-N 4-(3,5-dimethylphenyl)-1,3-thiazol-2-amine Chemical compound CC1=CC(C)=CC(C=2N=C(N)SC=2)=C1 MGWGWNFMUOTEHG-UHFFFAOYSA-N 0.000 description 4
- CBENFWSGALASAD-UHFFFAOYSA-N Ozone Chemical compound [O-][O+]=O CBENFWSGALASAD-UHFFFAOYSA-N 0.000 description 4
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 description 3
- 229910002651 NO3 Chemical group 0.000 description 3
- NHNBFGGVMKEFGY-UHFFFAOYSA-N Nitrate Chemical group [O-][N+]([O-])=O NHNBFGGVMKEFGY-UHFFFAOYSA-N 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 229910002091 carbon monoxide Inorganic materials 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 230000001105 regulatory effect Effects 0.000 description 3
- 150000003863 ammonium salts Chemical class 0.000 description 2
- 230000000052 comparative effect Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 239000000446 fuel Substances 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 238000010206 sensitivity analysis Methods 0.000 description 2
- QAOWNCQODCNURD-UHFFFAOYSA-L sulfate group Chemical group S(=O)(=O)([O-])[O-] QAOWNCQODCNURD-UHFFFAOYSA-L 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 229910021529 ammonia Inorganic materials 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 102000030592 phosphoserine aminotransferase Human genes 0.000 description 1
- 108010088694 phosphoserine aminotransferase Proteins 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C10/00—Computational theoretical chemistry, i.e. ICT specially adapted for theoretical aspects of quantum chemistry, molecular mechanics, molecular dynamics or the like
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/10—Analysis or design of chemical reactions, syntheses or processes
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/20—Identification of molecular entities, parts thereof or of chemical compositions
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Chemical & Material Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Crystallography & Structural Chemistry (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Spectroscopy & Molecular Physics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Analytical Chemistry (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a method, a system and equipment for updating discharge list feedback based on a response surface model, wherein the method comprises the following steps: dividing the target area into a plurality of subareas; constructing a control matrix for the whole region and each sub-region; simulating a control scenario in the control matrix; calculating polynomial fitting equation coefficients of the single and integral regional response surfaces; calculating the weight of the overall change prediction error of the region; constructing a response surface model of the emission of the precursor to the target pollutant concentration; obtaining a plurality of groups of sampling results; acquiring an original list of atmospheric pollutants and air quality monitoring data of a target area; solving the optimal emission proportion of each pollutant in the sampling result; and constructing a loss function, and calculating a loss function value of each optimized result, wherein the minimum loss function value is used as a final correction result of the original list of the air pollutants. The invention supports real-time updating of the list and can correct PM once 2.5 、NH 3 And the pollutant emission amount of VOC and the like which are not easy to obtain monitoring data.
Description
Technical Field
The invention relates to the technical field of air quality monitoring, in particular to a method, a system and equipment for updating discharge list feedback based on a response surface model.
Background
The atmospheric pollutant emission list is necessary input data of air quality mode simulation, and is an important basis for air quality prediction and emission reduction policy evaluation. However, the current mainstream "bottom-up" method of creating the atmospheric pollutant emission list requires detailed statistics of a large amount of emission data, so that the newly created emission list is usually one year or more later than the current situation, and the actual emission amount is obviously changed internationally due to the update of the emission reduction policy and technology, so that the error of the "bottom-up" updated list is very large, and the air quality forecast and the emission reduction policy formulation are seriously affected.
The current correction or calibration methods for the atmospheric pollutant emission list are mainly based on air quality monitoring data to adjust the list, and the specific methods mainly comprise three types: (1) A numerical source analysis method of an air quality mode is used for obtaining the contribution quantity of a pollution source and correcting the emission quantity; (2) Establishing a source-receptor relationship by adopting a linear model, and solving a correction value of the discharge amount by using a linear programming; (3) The air quality mode is added into the calculation process of the modern optimization algorithm, and the emission of the optimization list is calculated through a large number of iterations.
As the current national air quality evaluation standard only comprises 6 pollutants (nitrogen dioxide, sulfur dioxide, carbon monoxide, ozone and PM 10 、PM 2.5 ) Most monitoring stations also monitor only these 6 pollutants, and the atmospheric pollutant emissions list includes NH in addition to 3 And VOC and the like, which are indispensable for the simulation of air quality. Due to the lack of monitoring data, the NH cannot be updated by the method 3 And VOC emissions, which are less accurate and practical.
In addition, for the method (1), the existing source analysis or sensitivity analysis methods (DDM, HDDM, ISAM, PSAT, OSAT) with built-in air quality modes have large uncertainty, and reliable emission lists are difficult to obtain by using source analysis or sensitivity analysis data; for the method (2), the pollutant discharged into the atmosphere can undergo a series of complex physicochemical processes, the source-receptor relationship of which is highly nonlinear, and a linear model can bring about larger errors; for the method (3), the source-receptor relationship of the method is the air quality mode, and the linear assumption problem in the method (2) is avoided, but the air quality mode simulation calculation cost is high, and the modern optimization algorithm needs a large amount of iterative calculation, so that the calculation power requirement of the method is extremely high, and the method is difficult to be suitable for large-scale problems such as list correction, air quality simulation and the like.
Disclosure of Invention
In order to overcome the defects and shortcomings in the prior art, the invention provides a response surface model-based emission list feedback updating method, which can accurately update an atmospheric pollutant emission list in real time and support NH (NH) for which monitoring data are difficult to obtain 3 And the emission quantity of pollutants such as VOC is corrected, the response curve model can calculate the concentration distribution of various pollutants under any emission scene in second-level time, the efficiency of updating the list is extremely high, the traditional annual updating of the list can be supported, the real-time updating of the list can be realized, and the performance of air quality simulation is greatly improved.
The second object of the invention is to provide an emission list feedback updating system based on a response surface model;
a third object of the present invention is to provide a computing device;
in order to achieve the above purpose, the present invention adopts the following technical scheme:
an emission list feedback updating method based on a response surface model comprises the following steps:
dividing the target area into a plurality of subareas based on the target area pollutant emission list adjustment task;
constructing a control matrix for the whole target area and each sub-area, and generating a plurality of control scenes for each area to form the control matrix;
Simulating control scenes in all control matrixes based on the air quality mode;
calculating coefficients of a polynomial fitting equation of a single area response surface and coefficients of a polynomial fitting equation of an integral area response surface according to results of air quality simulation of each scene and a control matrix;
calculating the sum of target pollutant contribution values of all areas to grid points based on the results of air quality simulation of each scene;
calculating to obtain a prediction error of the grid point caused by the overall change of the target area based on the coefficient of the polynomial fitting equation of the overall area response surface and the polynomial fitting equation of the overall area response surface;
calculating the weight of the overall change prediction error of the region according to the control matrix;
constructing a response surface model of precursor emission to target pollutant concentration based on the sum of target pollutant contribution values of all regions to a certain subarea of the target region, prediction errors, prediction error weights and target pollutant reference scene concentration values of grid points;
randomly sampling control factors of all areas to obtain a plurality of groups of sampling results as initial solutions of heuristic optimization algorithms;
acquiring an original list of atmospheric pollutants and air quality monitoring data of a target area;
Solving the optimal emission proportion of each pollutant in the sampling result based on a heuristic random search algorithm, and taking the optimal emission proportion as an optimization result of the sampling result;
constructing a loss function, calculating a loss function value corresponding to the optimized result based on the loss function, and selecting the optimized result with the minimum loss function value as a final correction result of the original list of the atmospheric pollutants.
As a preferable technical scheme, a control matrix is constructed for an integral target area and each sub-area, each area adopts a Hamersley sequence or Latin hypercube sampling mode to generate a control matrix composed of 20-40 control scenes, the integral target area control scene comprises 1 reference scene and 1 artificial source emission 100% reduction scene, and the control matrix of each sub-area comprises 1 artificial source emission 100% reduction scene.
As a preferable technical solution, the response surface model is expressed as:
ΔConc s→t ＝X s→t ×R s
wherein P is t Target pollutant concentration prediction value, ΔConc, representing grid point t All→t Representing the sum of target pollutant contribution values of all regions to region t, base t Target contaminant reference scene concentration value, ΔConc, representing grid point t s→t Representing the target pollutant contribution value of the region s to the grid point t, wherein the region s is a certain subarea of the target region, R s Representing a polynomial fit equation for a single-region response surface, X s→t Representing a fitting equation R with s as the source region and t as the target grid point s Is used for the coefficient of (a),representing the prediction error for grid point t due to the global change of the target area, +.>Polynomial fitting equation representing response surface for whole area, ++>Representing the whole area as the source area and t asFitting equation of target grid point->ω represents the weight of the region-wide variation prediction error.
As a preferable technical scheme, a polynomial fitting equation for a single-region response surface is represented by a vector composed of s-region control factors, when the target pollutant is PM 2.5 Or PM 2.5 In the form of R s，PM2.5 When the target pollutant is O 3 When the equation is in the form of R s，O3 The method is specifically expressed as follows:
R s，PM2.5 ＝[E NOx E NOx 2 E NOx 3 E SO2 E SO2 2 E NH3 E NH3 2 E VOC E NOx 2 E SO2 E NOx 2 E NH3 E NOx 2 E VOC E NOx E VOC E PM2.5 ]
R s，O3 ＝[E NOx E NOx 2 E NOx 3 E VOC E NOx E VOC E NOx 2 E VOC E SO2 E NH3 ]
wherein E is NOx Indicating emissions as NO x Emission single zone control factor, E SO2 Indicating emissions as SO 2 Emission single zone control factor, E NH3 Indicating the emission as NH 3 Emission single zone control factor, E VOC A single emission zone control factor, E, representing emissions as VOCs PM2.5 Indicating emissions as PM 2.5 The emission amount single region control factor of (c), where the single region means the s region.
As a preferable technical scheme, a polynomial fitting equation for the whole area response curve is represented by a vector composed of whole area control factors, when the target pollutant is PM 2.5 Or PM 2.5 In the form of the equationWhen the target pollutant is O 3 The equation form is->The concrete steps are as follows:
wherein,indicating emissions as NO x Is a total emission area control factor, +.>Indicating emissions as SO 2 Is a total emission area control factor, +.>Indicating the emission as NH 3 Is a total emission area control factor, +.>Indicating that emissions are an integral area control factor of the amount of emissions of VOCs.
As a preferable technical scheme, the weight of the prediction error of the overall change of the area is calculated according to the control matrix, specifically expressed as:
wherein E is NOx,s NO representing the s-region of the control matrix x Reducing the proportion E SO2,s SO representing the s-region in the control matrix 2 Reducing the proportion E NH3,s NH representing s-region in control matrix 3 Reducing the proportion;
the target pollutant is PM 2.5 And O 3 The weight of the overall change prediction error for the region is set to 1.
As an optimal technical scheme, solving the optimal emission proportion of each pollutant in the sampling result based on a heuristic random search algorithm comprises the following specific steps:
Calculating concentration distribution of the atmospheric pollutants in the target area corresponding to the initial sampling scene based on the response curved surface model;
calculating the ratio gamma of the concentration of the pollutant to the monitored concentration obtained by the response surface model in the same grid point sp,i The method is specifically expressed as follows:
wherein, monitor sp,i Representing the concentration monitoring value, model, of the species sp at the i-grid sp,i Representing the simulation value of the concentration of the response surface of the species sp at the i grid, gamma sp,i Representing the ratio of a monitoring value to a simulation value of the concentration of a species sp at the i grid, wherein the species sp refers to a target pollutant of the response surface model;
calculating an arithmetic average of the ratio in each sub-region of the target region, specifically expressed as:
wherein, gamma sp,s The average value of the ratio of the monitoring value to the analog value of each monitoring point in the region s of the species sp is represented, n represents the number of the monitoring points in the region s, and the region s is a certain sub-region of the target region;
calculating the maximum value of the absolute value of the species sp in all areas, expressed in detail as:
M sp ＝Max(|γ sp,s |)
wherein M is sp Representing the maximum value of the absolute value of the species sp in all regions;
judging whether the difference between the obtained pollutant concentration and the monitored concentration is within a preset threshold value range or not based on the maximum value of the absolute value of the species sp in all areas, if the difference is not within the threshold value range, adjusting the pollutant discharge proportion of the discharge scene according to a preset step length, re-using the response curved surface model to simulate the response concentration of the discharge proportion after the adjustment, repeating iteration until the error between the response concentration and the monitored concentration is smaller than a set value or the circulation reaches a set upper limit, selecting the discharge proportion with the minimum error as the optimal discharge proportion of the corresponding pollutant, and according to SO (solid oxide semiconductor) 2 、NH 3 、NO x VOC, primary PM 2.5 The discharge proportion of the precursor pollutants is corrected one by one, and the corresponding relation of the corresponding precursor pollutant discharge proportion is corrected one by one according to the difference between the simulated concentration and the monitored concentration of the target pollutants, namely (1) SO 2 -SO 2 ；(2)NH 3 -NH 3 ；(3)NO x -NO 2 ；(4)VOC-O 3 The method comprises the steps of carrying out a first treatment on the surface of the (5) Primary PM 2.5 Inhalable particulate PM 2.5 Wherein the former is a precursor contaminant and the latter is a corresponding target contaminant.
As a preferred technical solution, a loss function is constructed, specifically expressed as:
wherein Loss represents a Loss function value, gamma sp,s The average value of the ratio of the monitoring value to the simulation value of each monitoring point in a region s of a species sp, which is the target pollutant of the response surface model, is represented, and the region s is a certain subarea of the target region.
In order to achieve the second object, the present invention adopts the following technical scheme:
an emissions list feedback update system based on a response surface model, comprising: the system comprises a region dividing module, a control matrix constructing module, a control scene simulating module, a response surface polynomial fitting equation coefficient calculating module, a target pollutant contribution calculating module, a prediction error weight calculating module, a response surface model constructing module, a sampling result acquiring module, a target region data acquiring module, a sampling result optimizing module, a loss function constructing module and a final correction result calculating module;
The region dividing module is used for dividing the target region into a plurality of subareas based on the target region pollutant emission list adjustment task;
the control matrix construction module is used for constructing a control matrix for the whole target area and each sub-area, and each area generates a plurality of control scenes to form the control matrix;
the control scenario simulation module is used for simulating control scenarios in all control matrixes based on the air quality mode;
the response surface polynomial fitting equation coefficient calculation module is used for calculating the coefficient of the polynomial fitting equation of the single area response surface and the coefficient of the polynomial fitting equation of the whole area response surface according to the results of air quality simulation of all scenes and the control matrix;
the target pollutant contribution value calculation module is used for calculating the sum of target pollutant contribution values of all areas to grid points based on the results of air quality simulation of each scene;
the prediction error calculation module is used for calculating and obtaining a prediction error caused by the overall change of the target area to the grid points based on the coefficient of the polynomial fitting equation of the overall area response surface and the polynomial fitting equation of the overall area response surface;
The prediction error weight calculation module is used for calculating the weight of the overall change prediction error of the region according to the control matrix;
the response surface model construction module is used for constructing a response surface model of the precursor discharged to the target pollutant concentration based on the sum of target pollutant contribution values of all areas to a certain subarea of the target area, the prediction error weight and the target pollutant reference scene concentration value of the grid point;
the sampling result acquisition module is used for randomly sampling control factors of all areas to acquire a plurality of groups of sampling results as initial solutions of heuristic optimization algorithms;
the target area data acquisition module is used for acquiring an original list of the atmospheric pollutants and air quality monitoring data of the target area;
the sampling result optimization module is used for solving the optimal emission proportion of each pollutant in the sampling result based on a heuristic random search algorithm and taking the optimal emission proportion as an optimization result of the sampling result;
the loss function construction module is used for constructing a loss function;
the final correction result calculation module is used for calculating a loss function value corresponding to the optimization result based on the loss function, and selecting the optimization result with the minimum loss function value as the final correction result of the original list of the atmospheric pollutants.
In order to achieve the third object, the present invention adopts the following technical scheme:
a computing device comprises a processor and a memory for storing a program executable by the processor, wherein the processor realizes the discharge list feedback updating method based on the response surface model when executing the program stored by the memory.
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) According to the invention, by constructing the high-performance polynomial response curved surface model, the calculation efficiency and accuracy of simulation are improved, the pollutant environment concentration distribution under any emission scene can be simulated in second-level time, and the average relative error between the pollutant environment concentration distribution and the simulation result of the air quality mode is about 1%.
(2) The invention builds a pollution source list feedback updating method based on a response surface model based on a polynomial response surface model and combining with an atmospheric physicochemical reaction rule, and solves the problem of one-time PM 2.5 、NH 3 And cleaning of contaminants such as VOCs without direct monitoring dataThe single updating problem enables the invention to carry out inversion updating on all main pollutants involved in inventory, can accurately update the atmospheric pollutant emission inventory in real time, and supports the primary PM (particulate matter) of which the monitoring data is difficult to obtain 2.5 、NH 3 And the amount of pollutant such as VOC.
(3) The polynomial response curved surface model and the heuristic search algorithm constructed by the invention have high calculation efficiency, can update annual and monthly emission lists according to the needs, and can even update the emission lists in real time according to real-time monitoring data, thereby greatly improving the performance of air quality simulation.
Drawings
Fig. 1 is a flow chart of an exhaust list feedback updating method based on a response surface model in embodiment 1;
FIG. 2 is a schematic diagram of a simulated regional division method according to embodiment 1;
fig. 3 is a schematic flowchart of the implementation of the heuristic optimization algorithm for optimizing each sampling result in embodiment 1;
FIG. 4 (a) is a graph showing the comparison of the error before and after correction of the pollutant emission list in the simulated region of 10 months 2017 in example 1;
FIG. 4 (b) is a schematic diagram showing the comparison of the pollutant emission list of the simulation area of 10 months in 2017 in example 1 before and after correction;
FIG. 5 (a) is a graph showing the comparison of the error before and after correction of the pollutant emission list in the simulation area of 2019, 9 and 19 in the present example 1;
FIG. 5 (b) is a schematic diagram showing the comparative before and after correction of the pollutant emission list in the simulation area of 2019, 9 and 19 days in example 1;
FIG. 6 (a) is a graph showing the comparison of errors before and after correction of the pollutant emission list in the 2017 simulated area in example 1;
FIG. 6 (b) is a comparative schematic diagram of the simulated area pollutant emission list of 2017 in example 1 before and after correction;
fig. 7 is a schematic calculation time diagram of the polynomial response surface model of the present embodiment 1 for 10 random verification scenarios;
fig. 8 is a schematic diagram of simulation errors of the polynomial response surface model of the present embodiment 1 for 10 random verification scenarios.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Example 1
As shown in fig. 1, the present embodiment provides a method for updating feedback of an emission list based on a response surface model, which includes the following steps:
s1: dividing a target area into a plurality of subareas according to the business requirement of the list adjustment;
as shown in fig. 2, taking the example of the task of adjusting the inventory of the simulated region 2017, the simulated region is divided into 6 areas, namely an a area, a B area, a C area, a D area, an E area and other areas;
S2: the control matrix is established for the whole target area and each sub-area, and the embodiment preferably establishes the control matrix for the whole simulation area and the 6 areas, wherein each area is provided with 5 control factors, namely NO x 、SO 2 、NH 3 VOC and Primary PM 2.5 Generating 20-40 control scenes by adopting a Hamersley sequence sampling mode in each region to form a control matrix, wherein the overall target region control scene comprises 1 reference scene and 1 artificial source emission 100% reduction scene, and the control matrix of each sub-region comprises 1 artificial source emission 100% reduction scene;
s3: the air quality mode is used for simulating the control scenes in all the control matrixes, and the embodiment preferably uses a third generation air quality mode (CMAQ) for simulating the control scenes in all the control matrixes;
s4: constructing a response surface model of precursor emission to the target pollutant concentration by using the control matrix and the corresponding air quality simulation result;
the embodiment constructs based on the atmospheric physicochemical reaction rules and the related experiencePolynomial equations were constructed for fitting to precursor contaminants (NO x ，SO 2 、NH 3 VOC, primary PM 2.5 ) With target pollutant (NO) x ，SO 2 、NH 3 、VOC、PM 2.5 、O 3 ) The response relation between the two is that the combination and the square degree of each item of the polynomial are determined, and the coefficient of each item can be directly calculated by using a linear regression mode (the derivation and the iterative calculation of nonlinear regression are avoided), so that the improvement of the calculation efficiency quality is realized. In addition, since the fitting equation constructed in the embodiment is an explicit equation, the number of terms of the equation is small and the number of times is small, so that the calculation complexity is low, the calculation time can be controlled in the second level, and the specific steps include:
S41: based on the results of air quality simulation (ΔConc for each scenario s→t ) Control matrix (R s，PM2.5 ) Calculating coefficients (R s ) And coefficients of a polynomial fit equation for an integral regional response surface
ΔConC All→t The calculation mode of (a) is shown in formulas (1) - (4):
ΔConc s→t ＝X s→t ×R s (2)
R s,PM2.5 ＝[E NOx E NOx 2 E NOx 3 E SO2 E SO2 2 E NH3 E NH3 2 E VOC E NOx 2 E SO2 E NOx 2 E NH3 E NOx 2 E VOC E NOx E VOC E PM2.5 ] (3)
R s，O3 ＝[E NOx E NOx 2 E NOx 3 E VOC E NOx E VOC E NOx 2 E VOC E SO2 E NH3 ] (4)
wherein E is NOx Indicating emissions as NO x Emission single zone control factor, E SO2 Indicating emissions as SO 2 Emission single zone control factor, E NH3 Indicating the emission as NH 3 Emission single zone control factor, E VOC A single emission zone control factor, E, representing emissions as VOCs PM2.5 Indicating emissions as PM 2.5 Is a single emission zone control factor of E NOx And E is SO2 For example, if E NOx =0.7, representing NO x Is adjusted to 70% of the original discharge amount, E SO2 =0, representing SO 2 The emission amount of (2) is reduced to 0.ΔConc s→t The target pollutant contribution value of the s region to the grid point t is represented, the s region is a certain sub-region of the target region, and the present embodiment divides the analog region into 6 regions of a ground, B ground, C ground, D ground, E ground and others, and s represents any one region of a ground, B ground, C ground, D ground, E ground and others. R is R s Representing a polynomial fit equation for a single-region response surface, the equation is defined by an s-region control factor (E NOx ，E SO2 ，E NH3 ，E VOC ,E PM2.5 ) And (5) a vector representation of the composition. When the target pollutant is PM 2.5 Or PM 2.5 When a precursor of (e.g. NO) x ，NO 2 ，NO，SO 2 ，NH 3 VOC), equation form R s,PM2.5 When the target pollutant is O 3 When the equation is in the form of R s,O3 ；X s→t Representing a fitting equation R with s as the source region and t as the target grid point s Is obtained by linear regression.
The calculation mode of (a) is shown in formulas (5) - (6):
wherein,representing the prediction error for grid point t due to the global change of the target area, +.>Polynomial fitting equation representing response surface for whole area controlled by whole area control factor +.> And (5) a vector representation of the composition. When the target pollutant is PM 2.5 Or PM 2.5 When a precursor of (e.g. NO) x ，NO 2 ，NO，SO 2 ，NH 3 VOC), equation form is ∈>When the target pollutant is O 3 The equation form is-> Fitting equation representing grid points with global area as source area and t as target>Coefficient of (2) by linear regressionObtaining the product.
S42: the weight omega of the prediction error is calculated according to the control matrix, and the calculation mode of omega is shown in formulas (7) - (8):
wherein E is NOx,s NO representing the s-region of the control matrix x Reducing the proportion E SO2,s SO representing the s-region in the control matrix 2 Reducing the proportion E NH3,s NH representing s-region in control matrix 3 Reducing the proportion;
s43: construction of a model of the response surface of precursor emissions to target contaminant concentrations (P t )。
Wherein P is t Target pollutant concentration prediction value, ΔConc, representing grid point t All→t Represents the sum of the target pollutant contribution values of all the regions to the t region, ω represents the weight of the prediction error of the overall change of the region, and the target pollutant is PM 2.5 And O 3 When the precursor of (e.g. NO x ，NO 2 ，NO，SO 2 ，NH 3 ，VOC)，ω＝1，Representing prediction error of grid point t due to overall change of target area, base t A target contaminant reference scene concentration value representing a grid point t;
s5: the control factors of all the areas are randomly sampled (including uniformly distributed sampling, latin hypercube sampling or Hamersley sequence sampling), and 10-200 groups of the total sampling are taken as initial solutions of heuristic optimization algorithms.
In this embodiment, since the dimensions are high for the 6 regions and 5 precursor contaminants involved, it is preferable to use 100 sets of evenly distributed samples as the initial solution for the heuristic optimization algorithm.
S6: preparing an original list of atmospheric pollutants and air quality monitoring data of a target area, wherein the monitoring data comprises longitude and latitude of a monitoring point and 6 items of atmospheric pollutants (nitrogen dioxide, sulfur dioxide, carbon monoxide, ozone and PM) 10 、PM 2.5 ) The pollutant concentration monitoring data can be an hour average value, a day average value, a month average value or an annual average value according to the compiling or correction requirements of the inventory (the hour average value and the day average value can be used for real-time feedback update of the inventory).
In this example, the statistics simulate 6 atmospheric pollutants (nitrogen dioxide, sulfur dioxide, carbon monoxide, ozone, PM) in the above 6 regions 10 、PM 2.5 ) Is a 10 month average concentration monitor.
S7: optimizing each sampling result in the step S5 by using a heuristic optimization algorithm, and solving the optimal emission proportion of each pollutant by using a heuristic random search algorithm on the basis of a response surface model, wherein the implementation process is as follows:
according to SO 2 、NH 3 、NO x VOC, primary PM 2.5 Sequentially correcting the emission of pollutant species by SO 2 For example, first randomly generate a number of eligible SOs 2 Discharge ratio, and simulating the SO by using a response surface model 2 The discharge ratio is corresponding to SO 2 Responsive to concentration, SO 2 Comparing the response concentration with the monitored concentration, if the response concentration of a certain emission scenario is larger than the monitored concentration, reducing the emission proportion of the emission scenario by a certain step (the default of the embodiment is 2 percent) (if the response concentration is smaller than the monitored concentration, increasing the emission proportion by a certain step), and adjusting the SO 2 Emission ratioExample re-use of the response surface model to simulate its response concentration, repeating the above-mentioned process until SO 2 After the error between the response concentration and the monitored concentration is less than the set value (default of 5% in this embodiment) or the cycle reaches the set upper limit, the SO with the smallest error is selected from the several results 2 Discharge ratio as SO 2 And uses that ratio for later species correction operations.
For VOC emissions, which are generally free of direct monitoring data, the algorithm is described in NO x After the emission ratio is determined, the VOC emission ratio is randomly generated and the corresponding O is simulated by using a curved surface model 3 Response concentration, using O 3 Monitoring concentration constraint to solve the optimal emission proportion of VOC;
in particular, for NH without direct monitoring data in general 3 Emissions, the present embodiment uses PM 2.5 PM calculation from component data of (a) 2.5 The ratio of the ammonium salt relative to the sulfate and nitrate, and thus the environmental NO is combined according to the ratio x And SO 2 Concentration estimation environment NH of (c) 3 And the ambient NH estimated by the method 3 Concentration constraint solving for NH 3 Is used for the fuel injection. After the optimal emission ratio of each pollutant is obtained, the optimal emission ratio is multiplied by the prior inventory emission amount to obtain a corrected posterior inventory.
The detailed process of the heuristic optimization algorithm is as follows:
s71: calculating 5 atmospheric pollutants (nitrogen dioxide, sulfur dioxide, ammonia, ozone and PM) corresponding to the initial sampling scene by using the response surface model of the step S4 2.5 ) Concentration distribution in the target area, i.e., the concentration distribution in the simulation area of the present embodiment;
S72: calculating the ratio gamma of the concentration of the pollutant to the monitored concentration obtained by the response surface model in the same grid point sp,i The method is specifically expressed as follows:
wherein, monitor sp,i Representing the position of the i gridMonitoring the concentration of sp, model sp,i Representing the simulation value of the concentration of the response surface of the species sp at the i grid, gamma sp,i Representing the ratio of the monitored value of the sp concentration of the species at the i-grid to the analog value;
in particular, if NH 3 Is missing, PM can be used 2.5 Component monitoring data and NO 2 、SO 2 Environmental concentration monitoring data, deducing NH according to proportion 3 See equation (12).
Wherein, monitor NH3,i 、Monitor NO2,i 、Monitor SO2,i Respectively represent NH 3 、NO 2 And SO 2 Ambient concentration, PM, at the i-grid NH4 ，PM NO3 ，PM SO4 Respectively represents ammonium salt, nitrate and sulfate in PM 2.5 The content of (3) is as follows.
S73: the ratio gamma of the pollutant concentration to the monitoring concentration obtained according to the calculated response curve model in the same grid point sp，i (step S72) an arithmetic mean of the ratio in each sub-region of the target region (i.e., the simulated region) is calculated, specifically expressed as:
M sp ＝Max(|γ sp,s |) (14)
wherein, gamma sp,s Mean value representing ratio of monitoring value to analog value of species sp in region s, the species sp being target pollutant of response surface model, namely SO 2 、NH 3 、NO 2 、O 3 And PM 2.5 I represents the grid of the monitoring points in the region s, n represents the number of the monitoring points in the region s, M sp Represents the maximum value of the average absolute error of the sp of the species in all regions, if gamma sp，s =1, seeFor no difference between the monitored value and the analog value, if gamma sp，s If=1.07, the monitored value is considered to be 7% greater than the analog value, i.e., the difference is 7%, if γ sp,s =0.97, the difference between the monitored value and the analog value is 3%. Similarly, M sp =0.99, regarding that the maximum value of the average absolute error between each area monitor value and the analog value is 1%, the error is acceptable within the threshold range (preferably 5%) of the present embodiment;
because of the very complex physical and chemical relationships between the atmospheric pollutants, adjusting the concentration of a pollutant emission can have an effect on the concentration of other atmospheric pollutants, for example: SO reduction 2 Upon discharge, SO in the atmosphere 2 And PM 2.5 The concentration is reduced, at O 3 To a certain extent reduce NO x NO in the atmosphere x (including NO 2 And NO) concentration will decrease, but O 3 The concentration will increase. Therefore, the conventional method is difficult to complete the update and calibration work of the emission list of all pollutants under the complex emission-concentration response relation, in order to solve the problem, according to the nonlinearity degree of the response relation of each target pollutant (the pollutant of which the emission amount needs to be regulated in the list) and the chemical reaction relation of each pollutant in the atmosphere, the calculation sequence shown in the following table 1 is constructed, and the emission condition of each pollutant is regulated according to the sequence, so that the influence of the emission amount regulation of the subsequent pollutant on the accuracy of the emission amount of the pollutant regulated before can be reduced to the greatest extent, and the improvement of efficiency and accuracy is realized.
And judging whether the difference between the concentration of the pollutants obtained by the corresponding response surface model and the monitored concentration is within an acceptable range, if the maximum value of the absolute values of the errors of the species in all areas is less than 5%, continuously judging whether the concentration error of the next pollutant is acceptable, if the errors of all the pollutants in all the subareas are acceptable, ending the list adjustment work, otherwise, executing step S74.
TABLE 1 calculation sequence table for concentration and monitoring concentration difference of response surface model
S74: to solve one-time PM 2.5 And VOC, respectively, and correcting the emission of the precursor one by one according to the sequence of the table 2 and the corresponding relation thereof, and can be used for PM 2.5 (Main precursor: NO) x ，SO 2 ，NH 3 VOC, primary PM 2.5 ) And O 3 (Main precursor: NO) x VOC) and using O based on a response surface model 3 And PM 2.5 Ambient concentration of (c) is respectively for VOC and primary PM 2.5 Updating and calibrating the discharge amount of the fuel; if gamma is sp,s If the emission ratio of the species in the corresponding area is reduced by 2 percent (default adjustment step length can be customized according to the requirement), if gamma is larger than 1 sp,s And if the emission ratio of the species in the corresponding area is less than 1, the emission ratio of the species in the corresponding area is increased by 2 percent (the default adjustment step length can be customized according to the requirement). Returning again to S71, until the errors of all contaminants in all regions are acceptable or the cycle reaches a set upper limit, the inventory adjustment is completed and the optimization result of the sampling is returned.
TABLE 2 precursor-target contaminant correspondence correction relationship table
S8: after all sampling results are optimized, calculating a loss function value corresponding to the optimized result by using a formula (15), and selecting the optimized result with the minimum loss function value as a final correction scheme of an original list of the atmospheric pollutants, wherein the final result is obtained as shown in fig. 4 (a) and 4 (b);
in the embodiment, taking a simulation area 2017 month 10 list adjustment task as an example (updating a list in month), the implementation process of updating a month discharge list is demonstrated, the input data is modified into a pollutant year discharge list to be corrected and year-average monitoring data of corresponding pollutants, so that the year-by-year update list can be realized, the input data is modified into real-time data, and the real-time update of the discharge list can be realized;
taking a simulation area pollutant emission list of 2019, 9 and 19 days as an example (a real-time update list), as shown in fig. 5 (a) and 5 (b), obtaining a real-time correction list before-after correction comparison and error comparison; taking a 2017 simulated regional pollutant emission list as an example (updated by year), as shown in fig. 6 (a) and 6 (b), obtaining a comparison before and after the annual correction list is corrected and an error comparison;
the loss function value of this embodiment is specifically expressed as:
Wherein Loss represents a Loss function value, gamma sp,s The average value of the ratio of the monitored value to the analog value of each monitored point in the region s of the species sp is represented.
The invention can accurately update the emission list of the atmospheric pollutants in real time and support the NH which is difficult to obtain the monitoring data 3 And the emission amount of pollutants such as VOC (volatile organic compounds) is corrected, as shown in fig. 7 and 8, and the pollutant emission list in the area is simulated in 10 months in 2017 is taken as an example, the response surface model can calculate the concentration distribution of various pollutants under any emission scene in second level time, can improve the efficiency and accuracy of updating the list, can support the traditional annual update of the list, can update the list in real time, and greatly improves the air quality simulation performance.
Example 2
This embodiment is the same as embodiment 1 except for the following technical matters;
the embodiment provides an emission list feedback updating system based on a response surface model, which comprises the following steps: the system comprises a region dividing module, a control matrix constructing module, a control scene simulating module, a response surface polynomial fitting equation coefficient calculating module, a target pollutant contribution calculating module, a prediction error weight calculating module, a response surface model constructing module, a sampling result acquiring module, a target region data acquiring module, a sampling result optimizing module, a loss function constructing module and a final correction result calculating module;
In this embodiment, the area dividing module is configured to divide the target area into a plurality of sub-areas based on the target area pollutant emission list adjustment task;
in this embodiment, the control matrix construction module is configured to construct a control matrix for the entire target area and each sub-area, and each area generates a plurality of control scenarios to form the control matrix;
in this embodiment, the control scenario simulation module is configured to simulate control scenarios in all control matrices based on an air quality mode;
in this embodiment, the coefficient calculation module of the polynomial fitting equation of the response surface is configured to calculate, according to the result of air quality simulation of each scenario and the control matrix, the coefficient of the polynomial fitting equation of the single area response surface and the coefficient of the polynomial fitting equation of the whole area response surface;
in this embodiment, the target pollutant contribution value calculation module is configured to calculate a sum of target pollutant contribution values of all areas to grid points based on a result of air quality simulation of each scenario;
in this embodiment, the prediction error calculation module is configured to calculate, based on a coefficient of a polynomial fitting equation of the overall area response surface and a polynomial fitting equation of the overall area response surface, a prediction error caused by overall change of the target area on the grid point;
In this embodiment, the prediction error weight calculation module is configured to calculate a weight of a prediction error of an overall change of a region according to a control matrix;
in this embodiment, the response surface model building module is configured to build a response surface model of the precursor emission to the target pollutant concentration based on the sum of the target pollutant contribution values of all the regions to a certain sub-region of the target region, the prediction error weight, and the target pollutant reference scene concentration value of the grid point;
in this embodiment, the sampling result obtaining module is configured to randomly sample control factors of all areas, and obtain a plurality of groups of sampling results as an initial solution of a heuristic optimization algorithm;
in this embodiment, the target area data acquisition module is configured to acquire an original list of air pollutants and air quality monitoring data of a target area;
in this embodiment, the sampling result optimization module is configured to solve an optimal emission ratio of each pollutant in the sampling result based on a heuristic random search algorithm, and use the optimal emission ratio as an optimization result of the sampling result;
in this embodiment, the loss function construction module is configured to construct a loss function;
in this embodiment, the final correction result calculation module is configured to calculate a loss function value corresponding to the optimization result based on the loss function, and select the optimization result with the smallest loss function value as the final correction result of the original air pollutant list.
Example 3
The present embodiment provides a computing device, which may be a desktop computer, a notebook computer, a smart phone, a PDA handheld terminal, a tablet computer, or other terminal devices with display functions, where the computing device includes a processor and a memory, where the memory stores one or more programs, and when the processor executes the programs stored in the memory, the method for updating the discharge list feedback based on the response surface model of embodiment 1 is implemented.
The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above examples, and any other changes, modifications, substitutions, combinations, and simplifications that do not depart from the spirit and principle of the present invention should be made in the equivalent manner, and the embodiments are included in the protection scope of the present invention.
Claims (10)
1. An emission list feedback updating method based on a response surface model is characterized by comprising the following steps:
dividing the target area into a plurality of subareas based on the target area pollutant emission list adjustment task;
constructing a control matrix for the whole target area and each sub-area, and generating a plurality of control scenes for each area to form the control matrix;
Simulating control scenes in all control matrixes based on the air quality mode;
calculating coefficients of a polynomial fitting equation of a single area response surface and coefficients of a polynomial fitting equation of an integral area response surface according to results of air quality simulation of each scene and a control matrix;
calculating the sum of target pollutant contribution values of all areas to grid points based on the results of air quality simulation of each scene;
calculating to obtain a prediction error of the grid point caused by the overall change of the target area based on the coefficient of the polynomial fitting equation of the overall area response surface and the polynomial fitting equation of the overall area response surface;
calculating the weight of the overall change prediction error of the region according to the control matrix;
constructing a response surface model of precursor emission to target pollutant concentration based on the sum of target pollutant contribution values of all regions to a certain subarea of the target region, prediction errors, prediction error weights and target pollutant reference scene concentration values of grid points;
randomly sampling control factors of all areas to obtain a plurality of groups of sampling results as initial solutions of heuristic optimization algorithms;
acquiring an original list of atmospheric pollutants and air quality monitoring data of a target area;
Solving the optimal emission proportion of each pollutant in the sampling result based on a heuristic random search algorithm, and taking the optimal emission proportion as an optimization result of the sampling result;
constructing a loss function, calculating a loss function value corresponding to the optimized result based on the loss function, and selecting the optimized result with the minimum loss function value as a final correction result of the original list of the atmospheric pollutants.
2. The method for updating the emission list feedback based on the response surface model according to claim 1, wherein a control matrix is constructed for an overall target area and each sub-area, 20-40 control scenes are generated by adopting a Hamersley sequence or Latin hypercube sampling mode for each area, the overall target area control scenes comprise 1 reference scene and 1 artificial source emission 100% reduction scene, and the control matrix of each sub-area comprises 1 artificial source emission 100% reduction scene.
3. The response surface model-based emissions inventory feedback updating method of claim 1, wherein the response surface model is expressed as:
ΔConc s→t ＝X s→t ×R s
wherein P is t Target pollutant concentration prediction value, ΔConc, representing grid point t All→t Representing the sum of target pollutant contribution values of all regions to region t, base t Target contaminant reference scene concentration value, ΔConc, representing grid point t s→t Representing the target pollutant contribution value of the region s to the grid point t, wherein the region s is a certain subarea of the target region, R s Representing a polynomial fit equation for a single-region response surface, X s→t Representing a fitting equation R with s as the source region and t as the target grid point s Is used for the coefficient of (a),representing the prediction error for grid point t due to the global change of the target area, +.>Polynomial fitting equation representing response surface for whole area, ++>Fitting equation representing grid points with global area as source area and t as target>ω represents the weight of the region-wide variation prediction error.
4. The response surface model-based emission inventory feedback updating method according to claim 3, wherein the polynomial fitting equation for a single-region response surface is represented by a vector composed of s-region control factors, when the target pollutant is PM 2.5 Or PM 2.5 In the form of R s,PM2.5 When the target pollutant is O 3 When the equation is in the form of R s,O3 The method is specifically expressed as follows:
R s,PM2.5 ＝[E NOx E Nox 2 E NOx 3 E SO2 E SO2 2 E NH3 E NH3 2 E VOC E NOx 2 E SO2 E NOx 2 E NH3 E NOx 2 E VOC E NOx E VOC E PM2.5 ]
R s，O3 ＝[E NOx E NOx 2 E NOx 3 E VOC E NOx E VOC E NOx 2 E VOC E SO2 E NH3 ]
wherein E is NOx Indicating emissions as NO x Emission single zone control factor, E SO2 Indicating emissions as SO 2 Emission single zone control factor, E NH3 Indicating the emission as NH 3 Emission single zone control factor, E VOC Single area control of emissions representing emissions as VOCsFactor of production, E PM2.5 Indicating emissions as PM 2.5 The emission amount single region control factor of (c), where the single region means the s region.
5. The response surface model-based emission inventory feedback updating method according to claim 3, wherein the polynomial fitting equation for the whole area response surface is represented by a vector composed of whole area control factors, when the target pollutant is PM 2.5 Or PM 2.5 In the form of the equationWhen the target pollutant is O 3 The equation form is->The concrete steps are as follows:
wherein,indicating emissions as NO x Is a total emission area control factor, +.>Indicating emissions as SO 2 Is a total emission area control factor, +.>Indicating the emission as NH 3 Is a total emission area control factor, +.>Indicating that emissions are an integral area control factor of the amount of emissions of VOCs.
6. The method for updating the feedback of the emission list based on the response surface model according to claim 3, wherein the weight of the prediction error of the overall change of the area is calculated according to the control matrix, specifically expressed as:
wherein E is NOx，s NO representing the s-region of the control matrix x Reducing the proportion E SO2，s SO representing the s-region in the control matrix 2 Reducing the proportion E NH3,s NH representing s-region in control matrix 3 Reducing the proportion;
the target pollutant is PM 2.5 And O 3 The weight of the overall change prediction error for the region is set to 1.
7. The method for updating the feedback of the emission list based on the response surface model according to claim 1, wherein the method for solving the optimal emission ratio of each pollutant in the sampling result based on the heuristic random search algorithm comprises the following specific steps:
calculating concentration distribution of the atmospheric pollutants in the target area corresponding to the initial sampling scene based on the response curved surface model;
calculating the concentration of the contaminant obtained by the response surface model in the same grid pointMonitoring the ratio gamma of the concentrations sp，i The method is specifically expressed as follows:
wherein, monitor sp，i Representing the concentration monitoring value of the species sp at the i-grid, modei sp，i Representing the simulation value of the concentration of the response surface of the species sp at the i grid, gamma sp，i Representing the ratio of a monitoring value to a simulation value of the concentration of a species sp at the i grid, wherein the species sp refers to a target pollutant of the response surface model;
calculating an arithmetic average of the ratio in each sub-region of the target region, specifically expressed as:
wherein, gamma sp，s The average value of the ratio of the monitoring value to the analog value of each monitoring point in the region s of the species sp is represented, n represents the number of the monitoring points in the region s, and the region s is a certain sub-region of the target region;
Calculating the maximum value of the absolute value of the species sp in all areas, expressed in detail as:
M sp ＝Max(|γ sp，s |)
wherein M is sp Representing the maximum value of the absolute value of the species sp in all regions;
judging whether the difference between the obtained pollutant concentration and the monitored concentration is within a preset threshold value range or not based on the maximum value of the absolute value of the species sp in all areas, if the difference is not within the threshold value range, adjusting the pollutant discharge proportion of the discharge scene according to a preset step length, re-using the response curved surface model to simulate the response concentration of the discharge proportion after the adjustment, repeating iteration until the error between the response concentration and the monitored concentration is smaller than a set value or the circulation reaches a set upper limit, selecting the discharge proportion with the minimum error as the optimal discharge proportion of the corresponding pollutant, and according to SO (solid oxide semiconductor) 2 、NH 3 、NO x VOC, primary PM 2.5 The emission proportion of the precursor pollutants is corrected one by one, and the corresponding relation of the emission proportion of the corresponding precursor pollutants is corrected one by one according to the difference between the simulated concentration and the monitored concentration of the target pollutants, wherein the relation is as follows: (1) SO (SO) 2 -SO 2 ；(2)NH 3 -NH 3 ；(3)NO x -NO 2 ；(4)VOC-O 3 The method comprises the steps of carrying out a first treatment on the surface of the (5) Primary PM 2.5 Inhalable particulate PM 2.5 Wherein the former is a precursor contaminant and the latter is a corresponding target contaminant.
8. The response surface model-based emission list feedback updating method according to claim 7, wherein the construction of the loss function is specifically expressed as:
Wherein Loss represents a Loss function value, gamma sp, The average value of the ratio of the monitoring value to the simulation value of each monitoring point in a region s of a species sp, which is the target pollutant of the response surface model, is represented, and the region s is a certain subarea of the target region.
9. An exhaust list feedback updating system based on a response surface model, comprising: the system comprises a region dividing module, a control matrix constructing module, a control scene simulating module, a response surface polynomial fitting equation coefficient calculating module, a target pollutant contribution calculating module, a prediction error weight calculating module, a response surface model constructing module, a sampling result acquiring module, a target region data acquiring module, a sampling result optimizing module, a loss function constructing module and a final correction result calculating module;
the region dividing module is used for dividing the target region into a plurality of subareas based on the target region pollutant emission list adjustment task;
the control matrix construction module is used for constructing a control matrix for the whole target area and each sub-area, and each area generates a plurality of control scenes to form the control matrix;
the control scenario simulation module is used for simulating control scenarios in all control matrixes based on the air quality mode;
The response surface polynomial fitting equation coefficient calculation module is used for calculating the coefficient of the polynomial fitting equation of the single area response surface and the coefficient of the polynomial fitting equation of the whole area response surface according to the results of air quality simulation of all scenes and the control matrix;
the target pollutant contribution value calculation module is used for calculating the sum of target pollutant contribution values of all areas to grid points based on the results of air quality simulation of each scene;
the prediction error calculation module is used for calculating and obtaining a prediction error caused by the overall change of the target area to the grid points based on the coefficient of the polynomial fitting equation of the overall area response surface and the polynomial fitting equation of the overall area response surface;
the prediction error weight calculation module is used for calculating the weight of the overall change prediction error of the region according to the control matrix;
the response surface model construction module is used for constructing a response surface model of the precursor discharged to the target pollutant concentration based on the sum of target pollutant contribution values of all areas to a certain subarea of the target area, the prediction error weight and the target pollutant reference scene concentration value of the grid point;
The sampling result acquisition module is used for randomly sampling control factors of all areas to acquire a plurality of groups of sampling results as initial solutions of heuristic optimization algorithms;
the target area data acquisition module is used for acquiring an original list of the atmospheric pollutants and air quality monitoring data of the target area;
the sampling result optimization module is used for solving the optimal emission proportion of each pollutant in the sampling result based on a heuristic random search algorithm and taking the optimal emission proportion as an optimization result of the sampling result;
the loss function construction module is used for constructing a loss function;
the final correction result calculation module is used for calculating a loss function value corresponding to the optimization result based on the loss function, and selecting the optimization result with the minimum loss function value as the final correction result of the original list of the atmospheric pollutants.
10. A computing device comprising a processor and a memory for storing a program executable by the processor, wherein the processor, when executing the program stored in the memory, implements the response surface model-based emissions list feedback updating method of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310781360.9A CN117059198B (en) | 2023-06-29 | 2023-06-29 | Emission list feedback updating method, system and equipment based on response surface model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310781360.9A CN117059198B (en) | 2023-06-29 | 2023-06-29 | Emission list feedback updating method, system and equipment based on response surface model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117059198A CN117059198A (en) | 2023-11-14 |
| CN117059198B true CN117059198B (en) | 2024-02-23 |
Family
ID=88657882
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310781360.9A Active CN117059198B (en) | 2023-06-29 | 2023-06-29 | Emission list feedback updating method, system and equipment based on response surface model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117059198B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111523717A (en) * | 2020-04-15 | 2020-08-11 | 北京工业大学 | Inversion estimation method for atmospheric pollutant emission list |
| CN114676601A (en) * | 2022-05-27 | 2022-06-28 | 湖南工商大学 | Emission cost calculation method and device, computer equipment and storage medium |
| CN115712981A (en) * | 2022-10-11 | 2023-02-24 | 重庆市生态环境科学研究院 | Method and system for analyzing fine particulate matter source based on receptor and chemical transmission model |
-
2023
- 2023-06-29 CN CN202310781360.9A patent/CN117059198B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111523717A (en) * | 2020-04-15 | 2020-08-11 | 北京工业大学 | Inversion estimation method for atmospheric pollutant emission list |
| CN114676601A (en) * | 2022-05-27 | 2022-06-28 | 湖南工商大学 | Emission cost calculation method and device, computer equipment and storage medium |
| CN115712981A (en) * | 2022-10-11 | 2023-02-24 | 重庆市生态环境科学研究院 | Method and system for analyzing fine particulate matter source based on receptor and chemical transmission model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117059198A (en) | 2023-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kwok et al. | Implementation and evaluation of PM2. 5 source contribution analysis in a photochemical model | |
| Meinshausen et al. | The shared socio-economic pathway (SSP) greenhouse gas concentrations and their extensions to 2500 | |
| Pendergrass et al. | Surface and top-of-atmosphere radiative feedback kernels for CESM-CAM5 | |
| Keller et al. | Application of random forest regression to the calculation of gas-phase chemistry within the GEOS-Chem chemistry model v10 | |
| Li et al. | Attribution of PM2. 5 exposure in Beijing–Tianjin–Hebei region to emissions: implication to control strategies | |
| JP2019090805A (en) | Processor for detection data of environmental sensor, method for processing detection data of environmental sensor, computer-readable storage medium, and environment sensor system | |
| Cherubini et al. | Global spatially explicit CO2 emission metrics for forest bioenergy | |
| CN112711893B (en) | Method and device for calculating contribution of pollution source to PM2.5 and electronic equipment | |
| CN114742460A (en) | Method and device for determining enterprise to be controlled, electronic equipment and storage medium | |
| CN112052619B (en) | Air pollution particle information optimization method and device and electronic equipment | |
| Hill et al. | The use of CO 2 flux time series for parameter and carbon stock estimation in carbon cycle research | |
| Rogers et al. | Improving representation of photosynthesis in Earth System Models | |
| CN114218751A (en) | Quantitative evaluation method and device for ozone pollution, computer equipment and storage medium | |
| Zhao et al. | A multiphase CMAQ version 5.0 adjoint | |
| CN114912079A (en) | Method, device and equipment for calculating carbon emission of sewage plant and storage medium | |
| CN117059198B (en) | Emission list feedback updating method, system and equipment based on response surface model | |
| Zammit-Mangion et al. | WOMBAT v1. 0: A fully Bayesian global flux-inversion framework | |
| Isukapalli et al. | Computationally efficient uncertainty propagation and reduction using the stochastic response surface method | |
| Nicholls et al. | Reduced complexity model intercomparison project phase 1: Protocol, results and initial observations | |
| Jahangiri et al. | The modeling of CO2 removal from a gas mixture by 2-amino-2-methyl-1-propanol (AMP) using the modified Kent Eisenberg model | |
| Osborne et al. | The missing aerosol response in twentieth-century mid-latitude precipitation observations | |
| CN112036038A (en) | Emission source emission amount adjusting and optimizing method, device, equipment and storage medium | |
| Müller et al. | LEAFC3-N photosynthesis, stomatal conductance, transpiration and energy balance model: Finite mesophyll conductance, drought stress, stomata ratio, optimized solution algorithms, and code | |
| CN116151469A (en) | Model for forecasting air quality | |
| CN108932483B (en) | Smell identification method and device based on machine olfaction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |